Image classification-in-idrisi-and-ilwis
- 1. Image Classification in IDRISI ANDES and ILWIS (PRACTICAL) Prepared by: Oluwafemi. A. OPALEYE For GPM III Nov, 2010
- 2. IDRISI ANDES Unsupervised Classification 1) Click Image Processing and navigate to Hard Classifier and to Cluster 2) Click the number of bands to be used and input them 3) Give name for the Output Image 4) Click Fine, under Generalization Rule 5) Set maximum number of clusters under the Clustering Rule 6) Click Ok to finish 7) To rename the classes; right click on individual class and click “update Legend” Supervised Classification 1) Click Display Launcher under Display 2) Browse for the Composite Image and click ok to display it 3) Click on Digitize tool from the tool bar 4) Give name for the layer to be created and check polygon under layer type. 5) Click Ok to start digitizing 6) Digitize the identified land feature after zooming and right click to finish. 7) Pick the Save Digitized Data tool to save 8) Pick the Digitize tool to add other features to the active layer. 9) Click Image Processing; navigate to Signature and MAKSIG (for signature extraction) 10) Input the vector file you created earlier 11) Click on Enter signature files names to name the Classes 12) Brose and input the bands 13) Click Ok to finish 14) Click Image Processing, navigate to Hard Classifier and then to MINDIST (for minimum distance classification) 15) Check infinite under Maximum Search Distance 16) Browse for the Signature Files named earlier 17) Give output file name 18) Click next, then OK to Finish
- 3. ILWIS Density Slicing 1) First Create a Group Domain 2) From File, navigate to Create then Domain 3) Under Type: check: Class and Group and click OK 4) From the Domain Group Menu; click Edit and navigate to Add Item 5) Give the Upper Bound value and Name. Note check from the band to be “Sliced” 6) Do same for the other Group/Class 7) Do perform the “Slicing” click on Slicing from the Operation List or through Operations-Image Processing – Slicing 8) Input the Raster Map 9) Give Name for the Output Raster Map 10) Under Domain; browse the Group Domain Created earlier and 11) Show to view the Raster Map Unsupervised Classification 1) Goto Operations, navigate to Image Processing and then Cluster 2) Click on the number of bands to be used and insert them accordingly 3) Insert the number of Cluster 4) Give Name for the Output Raster Map 5) Click Show to view the result. Supervised Classification (Maximum Likelihood) 1) Create Map List (File – Create – Map List), 2) Give name for Map List 3) Open the Map List (as Colour Composite) 4) Create Sample Sets (File – Create – Sample Set) 5) Specify name and also create a Domain 6) Choose ‘Map List’ create Maplist 7) Click OK 8) From the Domain Class Menu; click Edit and navigate to Add Item 9) Add the number of classes to be created 10) To create Sample Sets, choose the Normal Tool to digitize the identified Sites on the image 11) Digitize as many as possible on the identified sites and click on exit editor when through with digitizing. 12) Click on Operations – Image Processing – Classify 13) Choose “Maximum Likelihood” 14) Give Name for the Output Raster Map 15) Click show to view the result